Add web research context adapter

.github/workflows/dispatch_shadow_signal.yml

@@ -11,7 +11,7 @@ on:
         description: "Bridge provider"
         required: false
         type: choice
-        default: "auto"
+        default: "codex"
         options:
           - auto
           - api
@@ -39,7 +39,7 @@ jobs:
       BRIDGE_REPOSITORY: ${{ vars.CODEX_AUDIT_BRIDGE_REPOSITORY || 'QuantStrategyLab/AIAuditBridge' }}
       BRIDGE_REF: ${{ vars.CODEX_AUDIT_BRIDGE_REF || 'main' }}
       BRIDGE_TASK: long_horizon_signal_shadow
-      BRIDGE_PROVIDER: ${{ github.event.inputs.provider || 'auto' }}
+      BRIDGE_PROVIDER: ${{ github.event.inputs.provider || 'codex' }}
       SOURCE_REF: ${{ github.event.inputs.source_ref || 'main' }}
       SHADOW_SIGNAL_LABEL: long-horizon-shadow
     steps:

scripts/build_context_bundle.py

@@ -2,6 +2,7 @@
 from __future__ import annotations
 
 import argparse
+import datetime as dt
 import json
 import sys
 from pathlib import Path
@@ -18,6 +19,7 @@
     normalize_symbols,
     write_context_bundle,
 )
+from research_signal_context_pipelines.research_context_adapter import ResearchContextAdapter  # noqa: E402
 from research_signal_context_pipelines.theme_universe import (  # noqa: E402
     build_theme_context,
     load_symbol_theme_exposure,
@@ -50,9 +52,13 @@ def main() -> int:
         default="data/output/context_bundle/latest_context_bundle.json",
         help="Output JSON context bundle path",
     )
+    parser.add_argument("--web-research-sources", help="Optional JSON config with whitelist and web research sources")
+    parser.add_argument("--web-research-timeout", type=float, default=10.0, help="Timeout in seconds for web research fetches")
+    parser.add_argument("--web-research-max-entries", type=int, default=8, help="Maximum web research entries to include")
     args = parser.parse_args()
 
     symbols = normalize_symbols(args.symbols)
+    generated_at = dt.datetime.now(dt.timezone.utc)
     theme_context = None
     if not args.no_theme_context:
         theme_taxonomy_path = Path(args.theme_taxonomy)
@@ -62,6 +68,14 @@ def main() -> int:
             exposures = load_symbol_theme_exposure(theme_exposure_path, known_theme_ids=themes)
             theme_context = build_theme_context(symbols=symbols, themes=themes, exposures=exposures)
 
+    web_research_context = None
+    if args.web_research_sources:
+        web_research_context = ResearchContextAdapter(
+            Path(args.web_research_sources),
+            timeout_seconds=args.web_research_timeout,
+            max_entries=args.web_research_max_entries,
+        ).build_context(pit_timestamp=generated_at)
+
     try:
         bundle = build_context_from_source(
             symbols=symbols,
@@ -70,7 +84,9 @@ def main() -> int:
             end_date=args.end_date,
             lookback_days=args.lookback_days,
             theme_context=theme_context,
+            web_research_context=web_research_context,
             allow_partial_downloads=not args.strict_downloads,
+            generated_at=generated_at,
         )
     except Exception as exc:
         if not args.allow_download_errors:
@@ -79,6 +95,8 @@ def main() -> int:
             symbols=symbols,
             error=f"{type(exc).__name__}: {exc}",
             theme_context=theme_context,
+            web_research_context=web_research_context,
+            generated_at=generated_at,
         )
     output_path = Path(args.output)
     write_context_bundle(bundle, output_path)

scripts/post_shadow_signal_request.py

@@ -172,7 +172,7 @@ def parse_args() -> argparse.Namespace:
     parser = argparse.ArgumentParser(description="Create or update a long-horizon shadow signal issue.")
     parser.add_argument("--repo", required=True)
     parser.add_argument("--source-ref", default="main")
-    parser.add_argument("--provider", default="auto")
+    parser.add_argument("--provider", default="codex")
     parser.add_argument("--bridge-repository", default="QuantStrategyLab/AIAuditBridge")
     parser.add_argument("--as-of-date")
     parser.add_argument("--context-file", help="Optional JSON context bundle to embed in the issue body")

src/research_signal_context_pipelines/context_bundle.py

@@ -276,6 +276,7 @@ def build_context_bundle(
     symbols: list[str],
     generated_at: dt.datetime | None = None,
     theme_context: Mapping[str, Any] | None = None,
+    web_research_context: Mapping[str, Any] | None = None,
 ) -> dict[str, Any]:
     contexts: dict[str, SymbolContext] = {}
     warnings: list[str] = []
@@ -337,6 +338,8 @@ def build_context_bundle(
     }
     if theme_context is not None:
         bundle["theme_context"] = dict(theme_context)
+    if web_research_context is not None:
+        bundle["web_research"] = dict(web_research_context)
     return bundle
 
 
@@ -347,6 +350,7 @@ def build_error_context_bundle(
     as_of_date: dt.date | None = None,
     generated_at: dt.datetime | None = None,
     theme_context: Mapping[str, Any] | None = None,
+    web_research_context: Mapping[str, Any] | None = None,
 ) -> dict[str, Any]:
     as_of = as_of_date or dt.date.today()
     timestamp = generated_at or dt.datetime.now(dt.timezone.utc)
@@ -375,6 +379,8 @@ def build_error_context_bundle(
     }
     if theme_context is not None:
         bundle["theme_context"] = dict(theme_context)
+    if web_research_context is not None:
+        bundle["web_research"] = dict(web_research_context)
     return bundle
 
 
@@ -395,7 +401,9 @@ def build_context_from_source(
     lookback_days: int = 420,
     fetch_fn: Callable[..., Mapping[str, Any]] | None = None,
     theme_context: Mapping[str, Any] | None = None,
+    web_research_context: Mapping[str, Any] | None = None,
     allow_partial_downloads: bool = False,
+    generated_at: dt.datetime | None = None,
 ) -> dict[str, Any]:
     end = parse_price_date(end_date) if end_date else dt.date.today()
     start = parse_price_date(start_date) if start_date else end - dt.timedelta(days=int(lookback_days))
@@ -409,7 +417,13 @@ def build_context_from_source(
             fetch_fn=fetch_fn,
             allow_partial=allow_partial_downloads,
         )
-    return build_context_bundle(rows, symbols=symbols, theme_context=theme_context)
+    return build_context_bundle(
+        rows,
+        symbols=symbols,
+        generated_at=generated_at,
+        theme_context=theme_context,
+        web_research_context=web_research_context,
+    )
 
 
 def write_context_bundle(bundle: Mapping[str, Any], path: Path) -> None: